Re: BUG #13657: Some kind of undetected deadlock between query and "startup process" on replica.

On Wed, Oct 21, 2015 at 4:42 AM, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
wrote:

> On Mon, Oct 5, 2015 at 5:47 PM, Maxim Boguk <maxim(dot)boguk(at)gmail(dot)com> wrote:
> > I would like to check 9.4.5 but I prefer wait for release and not build
> it
> > by hands, and from my experience building from git tree could be tricky
> as I
> > don't have root access to install additional packages on the server.
> > However, I will try it in next few days (if I get an permission).
>
> Maxim, did you get the occasion to test your setup with 9.4.5? It
> seems that your environment is more likely to reproduce this deadlock.

​Sorry for late response, after upgrade to 9.4.5 this issue never happened
anymore (zero occurrences during the last week).
Before all cases of deadlock happened during wal replay of 2 specific
tables (and only them) of the database of few thousands. These tables are
somewhat special in sense they have medium insert/delete load but most time
they have zero alive tuples and size changes from 0 block to 1-2 block and
back quite a lot.​

​ I suppose deadlock had been happened when replay process replaying
truncation to zero pages after vacuum on the master db (but it's mostly
theory with no good evidence).
​

--
Maxim Boguk
Senior Postgresql DBA
http://www.postgresql-consulting.ru/ <http://www.postgresql-consulting.com/>

Phone RU: +7 910 405 4718
Phone AU: +61 45 218 5678

LinkedIn: http://www.linkedin.com/pub/maksym-boguk/80/b99/b1b
Skype: maxim.boguk
Jabber: maxim(dot)boguk(at)gmail(dot)com

"People problems are solved with people.
If people cannot solve the problem, try technology.
People will then wish they'd listened at the first stage."